CN107258581B - Monitoring method of pig's movement based on bluetooth ear tag - Google Patents

Abstract

The invention discloses a pig exercise amount monitoring system based on a Bluetooth ear tag, which is characterized by comprising the Bluetooth ear tag, a camera, a Bluetooth host, a switch, a server, a client and a PC (personal computer), wherein the Bluetooth ear tag is hung on a pig ear and used for acquiring pig exercise amount data and uploading the pig exercise amount data to the server through the Bluetooth host via the switch; the camera is used for collecting the motion state of the pig and uploading the motion state to the server through the switch; the server is respectively connected with the client and the PC and used for sending the motion amount data. The system monitors the motion state of the pig in real time for 24 hours without manual watching, analyzes and records the motion state of the pig, can find abnormal conditions in time and take corresponding measures, reduces the loss of a breeding plant caused by the abnormal conditions, and effectively improves the working efficiency of breeding personnel.

Description

Monitoring method of pig's movement based on bluetooth ear tag

technical field

The invention relates to the research fields of pig welfare breeding, BLE4.0 technology, sensor technology and algorithm, in particular to a system and method for monitoring the movement of pigs based on Bluetooth ear tags.

Background technique

With the increasing scale and intensification of the modern breeding industry, automated pig raising will become the development trend of the pig industry. The sow is the foundation of the pig farm. Farmers now pay more and more attention to the health of sows, and the amount of exercise reflects the physical health of the sows to a certain extent; the amount of exercise of boars is inseparably related to semen vitality, which directly affects the conception rate. Real-time and accurate monitoring of the amount of exercise of pigs, analysis of pig behavior and taking corresponding measures, can improve the health of sows, enhance the quality of boar semen, reduce piglet mortality, improve pig breeding welfare, reduce labor costs, and improve breeding. The level of automation is significant.

At present, BLE4.0 technology has been widely used in medical, smart home, automobile and other fields, but there are few automation equipment in the field of agricultural production and facility breeding industry. At present, the judgement of pig exercise in the breeding industry mainly relies on the long-term visual observation of the breeder, which is not only heavy work but also easily affected by the subjective experience of the breeder, so that the facility welfare breeding has not shown great practical benefits. Therefore, real-time and accurate monitoring of pig exercise is one of the inevitable development trends in the study of fine pig farming in the future.

SUMMARY OF THE INVENTION

Based on the problems existing in the background technology, the present invention proposes a pig movement monitoring system based on Bluetooth ear tags, which includes a Bluetooth ear tag, a camera, a Bluetooth host, a switch, a server, a client and a PC. It is used to collect the pig's movement data on the pig's ear, and upload it to the server through the switch through the Bluetooth host; the camera is used to collect the pig's movement status and upload it to the server through the switch; the server is connected to the client and the PC respectively for sending exercise data.

Preferably, the Bluetooth ear tag includes a circuit structure and a shell structure, and the circuit structure is arranged in the shell structure in the form of a PCB board.

Specifically, the circuit structure of the Bluetooth ear tag includes a button battery, a filter circuit, a Bluetooth chip, a radio frequency circuit and an accelerometer, the button battery is connected to the filter circuit, and the output end of the filter circuit is respectively connected to the Bluetooth chip and the accelerometer, and the Bluetooth chip and the accelerometer are respectively connected. The accelerometers are connected to each other, and the Bluetooth chip and the radio frequency circuit are connected to each other.

Preferably, the bluetooth chip is connected to the bluetooth host, the bluetooth host is connected to the server through the switch, and the server sends the exercise quantity data to the client and the PC.

Specifically, the shell structure of the Bluetooth ear tag includes a platform, a through hole and a circular groove, and the circular groove is arranged on the platform in a concave shape; a through hole is opened on the circular groove, and the through hole has an internal thread and is fixed with the PCB board. holes to match.

The invention also discloses a method for monitoring the movement of pigs based on bluetooth ear tags, which records the movement state video of each pig through a camera as a judgment basis for artificially calibrated pig movement state identification; collects the acceleration of pigs through bluetooth ear tags Data, the acceleration data is transmitted to the host in the form of broadcasting through Bluetooth technology, and uploaded to the server for storage to the database; by selecting the pig's motion state in the status video and the acceleration data collected by the Bluetooth ear tag at this time, it is analyzed by constructing a decision tree. , traverse each segmentation method of each variable, find the best segmentation point, and identify the motion state of the pig, which includes walking, running, jumping, and stationary; the system calculates the pig according to the motion state of the pig. Only exercise and make decisions, if any abnormal situation occurs, it will send out an abnormal warning of pigs to the PC client in time and send an alarm signal to the breeder's mobile phone.

The specific process is:

Step 1: Design a Bluetooth ear tag to collect acceleration data and preprocess the weighted average of the data, and send the processed acceleration data to the air in the form of broadcast;

Step 2: Design the Bluetooth receiving host module, and use the radio frequency circuit to capture the broadcast acceleration data packets in the air to prepare for the server to obtain data;

Step 3: The bluetooth host will upload all the obtained data to the server by connecting with the serial port server and use http communication and store the data in the database;

Step 4: Calculate the combined acceleration and acceleration difference in the server, and add the original corresponding three-axis acceleration data as a complete data record to prepare for subsequent data analysis and feature extraction;

Step 5: Using the decision tree analysis method, the record (a _xi , a _yi , a _zi , a _i , Δa _i ) is regarded as a node, and all nodes are divided into training nodes and test nodes. Go to the video to perform manual calibration to determine the movement state of the pig (walking, running, jumping, stationary), and adopt the "learning with tutor" method for high-intensity sample training;

来量化它的“纯度”，直到每个节点足够“纯”为止；Step 6: Through the training nodes calibrated in Step 5, traverse each segmentation method of each variable to find the best segmentation point, and divide it into 4 nodes in total: walking, running, jumping, still, for 4 Nodes perform steps 2 to 3 respectively, using entropy

to quantify its "purity" until each node is sufficiently "pure";

Step 7: Determine the test nodes through the trained algorithm, and determine the pros and cons of the algorithm and algorithm parameters by calculating the ratio between the success and failure of the test;

Step 8: If the test result is poor, the operation process between steps 5 and 7 will be repeated until the accuracy is satisfied;

Step 9: Give each motion state a score by customizing the quantification standard;

Step 10: Accumulate the points of the exercise volume at a fixed time interval to calculate the exercise volume and the degree of exercise of the pigs. If there is an abnormal situation, send an abnormal warning to the PC client and send an alarm signal to the breeder's mobile phone.

Specifically, in the fourth step, the acceleration data needs to undergo the following preprocessing:

The collected acceleration data is filtered and denoised, and the data is preprocessed by the weighted average filtering algorithm:

Perform weighted average filtering and denoising on 100 sets of X-axis, Y-axis, and Z-axis acceleration data collected every 1 second to obtain preprocessed acceleration data, and use the following formula to calculate the average value:

In the formula, a ₀ , a ₁ ,...,a _N are all constants and should satisfy the following formula:

计算合加速度a，其中a_x，a_y，a_z分别表示X轴，Y 轴，Z轴方向的加速度pass-through

Calculate the resultant acceleration a, where a _x , a _y , and a _z represent the acceleration in the X-axis, Y-axis, and Z-axis directions, respectively

其中i表示第i秒。Calculate for each set of data

where i represents the ith second.

Preferably, the Bluetooth ear tag collects 100 sets of acceleration data within 1s. Since the data at a certain time cannot determine what kind of movement state the pig is in, it is necessary to determine the movement state of the pig once within a certain period of time. In the fifth step, each time 10 seconds to judge the movement state of the pig.

其中y_n表示10秒内运动量量化标准所对应的总分数，S_n表示不同运动状态所对应的分数，t_n为运动行为所对应的持续时间。Specifically, in the tenth step, an interval of 10 seconds is used as a unit to calculate the amount of exercise.

Among them, y _n represents the total score corresponding to the quantification standard of the amount of exercise in 10 seconds, _Sn represents the score corresponding to different exercise states, and t _n is the duration corresponding to the exercise behavior.

The beneficial effects of the present invention

(1) The system conducts 24-hour real-time monitoring of the movement status of pigs without manual duty. The system analyzes and records the movement status of pigs, and can detect abnormal conditions in time and take corresponding measures to reduce the losses caused by abnormal conditions in the breeding plant. Effective Improve the work efficiency of farmers.

(2) Compared with the traditional pig detection system, the system has low cost and convenient installation. At the same time, it provides a scientific basis for realizing welfare and automation of feeding management.

(3) The system adopts Bluetooth 4.0 wireless communication technology, which avoids the shortage of cost and stability caused by traditional wiring, and the system can ensure long-term stable work. In addition, the wireless technology facilitates the real-time collection of data and ensures the working efficiency of the system.

Description of drawings

FIG. 1 is a system block diagram of the present invention.

FIG. 2 is a circuit block diagram of the Bluetooth ear tag of the present invention.

FIG. 3 is an overall block diagram of the present invention.

FIG. 4 is a structural diagram of the shell of the Bluetooth ear tag of the present invention.

FIG. 5 is a circuit diagram of a power module.

Figure 6 is an accelerometer circuit diagram.

Figure 7 is a circuit diagram of the status indicator light.

Figure 8 is a radio frequency circuit diagram.

Detailed ways

Below in conjunction with embodiment, the present invention is further described, but protection scope of the present invention is not limited to this:

Embodiment 1: in conjunction with Fig. 1, a pig movement monitoring system based on bluetooth ear tag, which includes bluetooth ear tag, camera, bluetooth host, switch, server, client and PC, and the bluetooth ear tag is hung on the pig's ear. The above is used to collect the pig's movement data, and upload it to the server through the switch through the Bluetooth host; the camera is used to collect the pig's movement state and upload it to the server through the switch; the server is connected to the client and the PC to send the movement data.

Embodiment 2: The monitoring system according to Embodiment 1, the Bluetooth ear tag includes a circuit structure and a casing structure, and the circuit structure is arranged in the casing structure in the form of a PCB board.

Embodiment 3: With reference to FIG. 2, in the monitoring system described in Embodiment 2, the circuit structure of the Bluetooth ear tag includes a button battery, a filter circuit, a Bluetooth chip, a radio frequency circuit and an accelerometer, the button battery is connected to the filter circuit, and the filter The output ends of the circuit are respectively connected to the Bluetooth chip and the accelerometer, the Bluetooth chip and the accelerometer are connected to each other, and the Bluetooth chip and the radio frequency circuit are connected to each other.

In some preferred embodiments:

Referring to Figure 5, the power supply module supplies power to the Bluetooth ear tag through the button battery, and a 2.2uf capacitor is connected in parallel at both ends of the power supply to reduce the AC internal resistance of the power supply and prevent the electronic circuit from producing parasitic oscillations and other adverse consequences.

Combined with Figure 6, a 4.7K resistor is connected in series with the SCL and SDA pins of the accelerometer, and its potential is pulled up through the VCC (3.3V) power supply to ensure that it can work normally, and it is also connected in parallel on both sides of the power supply A 100nf capacitor is used for a certain degree of filtering.

Combined with Figure 7, a status indicator is added to the PCB circuit design, which is controlled by the P1.0 pin of the main chip, and a 470 ohm resistor is connected in series to form an indicator loop. When the Bluetooth ear tag is installed with a button battery, it works normally. , the status indicator will flash 3 times at an interval of 500ms, reminding the user that the Bluetooth ear tag is in a normal operation state, and then the status indicator will no longer flash, reducing the power consumption of the Bluetooth ear tag.

With reference to Figure 8, the 2450BM15A002 balun filter is used in the RF circuit to carry the on-board RF antenna AN1 for wireless RF signal transmission and reception, so as to prepare for the normal reception between the Bluetooth ear tag and the Bluetooth host.

The measured data is shown in Table 1 when the sending power of the slave is 0db and the distance between the sending and receiving modules is 15 meters.

Table 1 Stability test of data transmission

The RF circuit uses a balun filter and a 2.4G onboard PCB antenna recommended by TI. In actual measurement, the host can receive the broadcast data from the slave within a distance of 23m (without any obstructions), and transmit it stably. When the sending power of the slave is 0db, the sending module sends fixed 20 bytes of data. The measured transmission distance is shown in Table 2.

Table 2 Transmission distance test of Bluetooth ear tags

Adopt low-power Bluetooth chip with core, enable low-power mode in the protocol stack, and automatically enter different low-power modes (PM1, PM2, PM3) through system scheduling according to different applications and working states; In the ear tag program design, the fluctuating data within the set threshold range is adaptively filtered to reduce the broadcast power consumption of Bluetooth; the data is packaged and compressed to reduce the power consumption caused by turning on and off the broadcast; the selection of the acceleration sensor adopts Low-power accelerometer.

Embodiment 4, in conjunction with Fig. 3, the monitoring system as described in Embodiment 3, the bluetooth chip is connected with the bluetooth host, the bluetooth host is connected to the server through the switch, and the server sends the exercise data to the client and the PC (the server-side application mainly relies on the data in the database). The information saved in real time displays the movement information of the pigs in the pig house in the form of a visual interface).

Embodiment 5: with reference to FIG. 4 , in the monitoring system described in Embodiment 2, the shell structure of the Bluetooth ear tag includes a platform 1, a through hole 2 and a circular groove 4, and the circular groove 4 is concavely arranged on the platform 1; A through hole 2 is opened on the circular groove 4, and the through hole 2 has an inner thread and is matched with the fixing hole of the PCB board.

Embodiment 6: As described in Embodiment 1, a method for monitoring the movement of pigs based on Bluetooth ear tags, which records the motion state video of each pig through a camera, as a judgment basis for artificially calibrated pig motion state identification; Collect the acceleration data of the pig, transmit the acceleration data to the host in the form of broadcasting through Bluetooth technology, and upload it to the server to store it in the database; analyze by selecting the movement state of the pig in the status video and the acceleration data collected by the Bluetooth ear tag at this time , by constructing a decision tree, traversing each segmentation method of each variable, finding the best segmentation point, and identifying the motion state of the pig, which includes walking, running, jumping, and stationary; The movement state calculates the amount of movement of the pig and makes a decision. In the event of an abnormal situation, it will send an abnormal warning of the pig to the PC client in time and send an alarm signal to the breeder's mobile phone.

The present invention proposes a self-defined standard, the content of the standard is to quantify the scores of walking, running, jumping, and stillness, as shown in Table 3:

Table 3 Quantitative standards for the movement state of pigs

Embodiment 7: the monitoring method as described in embodiment 6, the concrete flow is:

Step 1: Design a Bluetooth ear tag to collect acceleration data and preprocess the weighted average of the data, and send the processed acceleration data to the air in the form of broadcast;

Step 2: Design the Bluetooth receiving host module, and use the radio frequency circuit to capture the broadcast acceleration data packets in the air to prepare for the server to obtain data;

Step 3: The bluetooth host will upload all the obtained data to the server by connecting with the serial port server and use http communication and store the data in the database;

Step 4: Calculate the combined acceleration and acceleration difference in the server, and add the original corresponding three-axis acceleration data as a complete data record to prepare for subsequent data analysis and feature extraction;

Step 5: Using the decision tree analysis method, the record (a _xi , a _yi , a _zi , a _i , Δa _i ) is regarded as a node, and all nodes are divided into training nodes and test nodes. Go to the video to perform manual calibration to determine the movement state of the pig (walking, running, jumping, stationary), and adopt the "learning with tutor" method for high-intensity sample training;

来量化它的“纯度”，直到每个节点足够“纯”为止；Step 6: Through the training nodes calibrated in Step 5, traverse each segmentation method of each variable to find the best segmentation point, and divide it into 4 nodes in total: walking, running, jumping, still, for 4 Nodes perform steps 2 to 3 respectively, using entropy

to quantify its "purity" until each node is sufficiently "pure";

Step 7: Determine the test nodes through the trained algorithm, and determine the pros and cons of the algorithm and algorithm parameters by calculating the ratio between the success and failure of the test;

Step 8: If the test result is poor, the operation process between steps 5 and 7 will be repeated until the accuracy is satisfied;

Step 9: Give each motion state a score by customizing the quantification standard;

Step 10: Accumulate the points of the exercise volume at a fixed time interval to calculate the exercise volume and the degree of exercise of the pigs. If there is an abnormal situation, send an abnormal warning to the PC client and send an alarm signal to the breeder's mobile phone.

Embodiment 8: The monitoring method as described in Embodiment 7, in the fourth step, the acceleration data needs to undergo the following preprocessing:

The collected acceleration data is filtered and denoised, and the data is preprocessed by the weighted average filtering algorithm:

Perform weighted average filtering and denoising on 100 sets of X-axis, Y-axis, and Z-axis acceleration data collected every 1 second to obtain preprocessed acceleration data, and use the following formula to calculate the average value:

In the formula, a ₀ , a ₁ ,...,a _N are all constants and should satisfy the following formula:

计算合加速度a，其中a_x，a_y，a_z分别表示X轴，Y 轴，Z轴方向的加速度pass-through

Calculate the resultant acceleration a, where a _x , a _y , and a _z represent the acceleration in the X-axis, Y-axis, and Z-axis directions, respectively

其中i表示第i秒。Calculate for each set of data

where i represents the ith second.

Embodiment 9: According to the monitoring method described in Embodiment 8, 100 sets of acceleration data are collected within 1 s of the Bluetooth ear tag. Since the data at a certain moment cannot determine the current state of movement of the pig, it must be determined once within a certain period of time. The movement state of the pig. In the fifth step, the movement state of the pig is judged every 10 seconds.

其中y_n表示10秒内运动量量化标准所对应的总分数，S_n表示不同运动状态所对应的分数，t_n为运动行为所对应的持续时间。Embodiment 10: The monitoring method as described in Embodiment 7, in the tenth step, taking the interval time of 10 seconds as a unit to calculate the amount of exercise,

Among them, y _n represents the total score corresponding to the quantification standard of the amount of exercise in 10 seconds, _Sn represents the score corresponding to different exercise states, and t _n is the duration corresponding to the exercise behavior.

The present invention also proposes a quantitative standard for the amount of exercise of pigs as shown in Table 4:

Table 4 Quantitative standard for the amount of exercise of pigs

Function description of the monitoring system proposed by the present invention:

(1) Collection and filtering of acceleration data

Because the pigs will occasionally flap their ears under normal activities, causing the Bluetooth ear tags to shake, resulting in abnormal acceleration data collected by the Bluetooth ear tags; the acceleration sensor itself also has certain zero drift and temperature drift, so it is necessary to The acceleration data is filtered to ensure the accuracy of the collected acceleration data.

(2) Transmission of data

The bluetooth chip transmits the collected acceleration data to the bluetooth host through the 2.4GHz ISM wireless frequency band technology through the bluetooth 4.0 technology, which provides a guarantee for the real-time collection of the pig's movement state data.

(3) Video capture

The video surveillance records the movement state of each pig. By selecting the specific movement state of the pig in the video and the acceleration data collected by the sensor at this time, the acceleration data is classified according to the movement state of the pig, and the corresponding relationship between the movement state of the pig and the acceleration data is determined. . Video surveillance is used as the basis for judging the motion status of pigs. After the corresponding relationship is established, the motion status of pigs only needs to be judged by real-time acceleration data, and there is no need for video collection.

(4) Analysis of the amount of exercise

The server establishes a motion state model through the received acceleration data, uses a specific algorithm to classify the motion state, and processes the data according to self-defined quantification standards, and finally stores it in the database, and sends out an abnormal situation to the PC client in time. The pigs are abnormally alerted and an alarm signal is sent to the breeder's mobile phone.

The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which the present invention pertains can make various modifications or additions to the described specific embodiments or substitute in similar manners, but will not deviate from the spirit of the present invention or go beyond the definitions of the appended claims range.

Claims (4)

1. a pig motion monitoring method based on bluetooth ear tag is characterized in that it records the motion state video of every pig by camera, as the judgment basis of artificial calibration pig motion state identification; Collect the acceleration of pig by bluetooth ear tag Data, the acceleration data is transmitted to the host through Bluetooth, and uploaded to the server for storage in the database; by selecting the pig's movement state in the status video and the acceleration data collected by the Bluetooth ear tag at this time, the movement state of the pig is identified, so The above motion states include walking, running, jumping, and standing still; the system calculates the amount of movement of the pigs and makes decisions based on the motion states of the pigs, and sends an abnormal warning to the PC client and the breeder’s mobile phone in time when an abnormal situation occurs. ; The specific process is: Step 1: Design a Bluetooth ear tag to collect acceleration data and preprocess the weighted average of the data, and send the processed acceleration data to the air in the form of broadcast; Step 2: Design the Bluetooth receiving host module, and use the radio frequency circuit to capture the broadcast acceleration data packets in the air to prepare for the server to obtain data; Step 3: The Bluetooth host will upload all the obtained data to the server and store the data in the database by connecting with the serial port server using http communication; Step 4: Calculate the combined acceleration and acceleration difference in the server, and add the original corresponding three-axis acceleration data as a complete data record to prepare for subsequent data analysis and feature extraction; Step 5: Using the decision tree analysis method, the record (a _xi , a _yi , a _zi , a _i , Δa _i ) is regarded as a node, and all nodes are divided into training nodes and test nodes. Go to the video to perform manual calibration to determine the movement state of the pig, and adopt the "learning with tutor" method for high-intensity sample training; Step 6: Through the training nodes calibrated in Step 5, traverse each segmentation method of each variable to find the best segmentation point, and divide it into 4 nodes in total: walking, running, jumping, still, for 4 The nodes perform the second to third steps respectively to calculate the entropy

Until the entropy of each node is higher than the set value; Step 7: Determine the test nodes through the trained algorithm, and determine the pros and cons of the algorithm and algorithm parameters by calculating the ratio between the success and failure of the test; Step 8: If the test result is poor, the operation process between steps 5 and 7 will be repeated until the accuracy is satisfied; Step 9: Give each movement state a score by customizing the quantification standard; Step 10: Accumulate the points of the exercise volume at a fixed time interval to calculate the exercise volume and the degree of exercise of the pigs. If there is an abnormal situation, send an abnormal warning to the PC client and send an alarm signal to the breeder's mobile phone. 2. monitoring method according to claim 1 is characterized in that in the 4th step, acceleration data needs to go through following preprocessing: The collected acceleration data is filtered and denoised, and the data is preprocessed by the weighted average filtering algorithm: Perform weighted average filtering and denoising on 100 sets of X-axis, Y-axis, and Z-axis acceleration data collected every 1 second to obtain preprocessed acceleration data, and use the following formula to calculate the average value:

In the formula, a ₀ , a ₁ ,...,a _N are all constants and should satisfy the following formula:

pass-through

Calculate the resultant acceleration a, where a _x , a _y , and a _z represent the acceleration in the X-axis, Y-axis, and Z-axis directions, respectively Calculate for each set of data

where i represents the ith second. 3. The monitoring method according to claim 1, wherein in the fifth step, the movement state of the pig is judged every 10 seconds. 4. monitoring method according to claim 1 is characterized in that in the tenth step, with the interval time of 10 seconds as a unit, carries out the calculation of the amount of exercise,

Among them, y _n represents the total score corresponding to the quantification standard of the amount of exercise in 10 seconds, _Sn represents the score corresponding to different exercise states, and t _n is the duration corresponding to the exercise behavior.

Images